Guides
SE Ranking API vs DataForSEO for AI search data in 2026
SE Ranking's AI Search endpoints cost €36 per 1,000 prompts read from a monthly database; DataForSEO runs your own prompt for $1.20 per 1,000 answers.

By Mathias C.SEO consultant, and more recently GEO, at MentionBox
These two keys answer different questions, which is why their rates look absurd side by side until you read what each one returns. SE Ranking's AI Search endpoints read a prompt database its documentation describes as "refreshed on a monthly basis", and bill 200 credits for every prompt they hand back, which works out at €36 per 1,000 against its wallet of €45 for 250,000 credits. DataForSEO's LLM Scraper runs the prompt you send on ChatGPT or Gemini and returns the answer, at $0.0012 per result on the standard queue, or $1.20 per 1,000. One tells you which prompts already mention a domain, the other tells you what an engine says today when you ask it something. API for GEO is our editorial pick when the deliverable is a computed metric per brand, after hands-on testing by Mathias C.
This face-off is written for the agency that reports AI visibility for several clients and has to choose a key, or budget for both. Below: which product on each side you are actually comparing, what one AI answer costs, the request you write, the fields that come back, the engines each key reaches, and the contract clauses that decide whether those numbers can go in something you sell. MentionBox, which publishes this site, has a commercial link with API for GEO.
Which SE Ranking API, and which DataForSEO API
Both names cover several products, and half the comparisons written about them put the wrong two in the same column. SE Ranking sells three routes to AI search data, DataForSEO two.
| Product | What it gives you | Whose prompts | Entry ticket |
|---|---|---|---|
| SE Ranking Data API, AI Search | metrics and prompts for any domain, read from a stored database | SE Ranking's, by country | €45 wallet, 250,000 credits |
| SE Ranking AI Results Tracker | daily checks on prompts you configure in a project | yours | Core subscription from €87.20 a month plus the API add-on |
| SE Visible API | SE Ranking's dedicated AI visibility platform, API marked beta | yours | $99 a month, Basic |
| DataForSEO AI Optimization API | the answer, its sources, brands, fan-out and shopping items | yours | $50 minimum payment |
| DataForSEO SERP API | Google AI Overviews and AI Mode with the sources each cites | yours | $50 minimum payment |
Published by each vendor on its pricing page and in its API reference, checked on October 23, 2026.
The pair that belongs in a head-to-head is the first line against the fourth, since those are the two self-serve keys you can buy today with no platform subscription behind them. They still differ on the one axis that decides everything downstream: SE Ranking's Data API answers questions about a domain, and DataForSEO's AI Optimization API answers questions you write.
If you want your own prompt list from SE Ranking, the route is the AI Results Tracker inside the Project API, which reads daily checks on prompts configured in a project and sits behind a Core, Growth or Enterprise subscription plus the API add-on at €39 a month billed annually. Its one genuinely good commercial detail: reading the answers costs nothing, the documentation stating that "retrieving prompt answers does not consume API credits or subscription quota". Our guide to tracking ten clients without seat fees covers why that line changes an agency's arithmetic.
The two AI search keys side by side
| API for GEO | SE Ranking Data API | DataForSEO AI Optimization | |
|---|---|---|---|
| What the key buys | one prompt run on one engine, enriched | reads over a stored prompt database | one prompt run on one engine, raw |
| Price per 1,000 AI answers | $13, down to $7 at volume | €36 on the wallet rate | $1.20 on the standard queue |
| Entry ticket | pay as you go, no minimum published | €45 wallet, 250,000 credits | $50 minimum payment |
| Engines, of the 10 we track | 10 | 5 | 5, plus 2 more on its SERP API key |
| Data freshness | on demand | "refreshed on a monthly basis" | on demand, up to 45 min standard |
| Share of voice | yes | yes, leaderboard endpoint | no |
| Sentiment | yes | no | no |
| Webhooks | yes | not documented | postback and pingback |
| SDKs | official SDKs | none published, Postman collection | Python, C#, TypeScript, Java |
| Rate limit | 300 requests a minute | 10 requests a second | up to 10,000 calls a minute |
| Resale | allowed | forbidden without written agreement | no clause published |
| Our score | 9.0 of 10 | 4.3 of 10 | 7.7 of 10 |
Published by each vendor on its pricing page and in its API reference, checked on October 23, 2026. For API for GEO, four of the ten engines, Gemini, Google AI Overviews, DeepSeek and Meta AI, plus the webhooks, the OpenAPI spec and the official SDKs, were confirmed in our hands-on test on October 26, 2026 and are not yet documented on apiforgeo.com.
Three rows carry most of the decision. The freshness row rules SE Ranking out of any weekly deliverable. The sentiment column is empty on both, so a client asking for a tone chart sends you to a third vendor whatever you pick here, a case our guide to DataForSEO alternatives with share of voice and sentiment works through. And the resale row is the one an agency reads last and should read first.
What one AI answer costs on each
SE Ranking prices by credit, and the credit cost per endpoint is where the bill actually forms. Its getting started page splits charges into a fixed cost "once per successful request" and a variable cost "for each record returned", and states that "only successful HTTP 2xx responses consume credits. 4xx and 5xx responses are not billed."
| SE Ranking AI Search endpoint | Credits | Cost on the €45 wallet |
|---|---|---|
GET /ai-search/discover-brand | 100 | €0.018 |
GET /ai-search/overview/by-engine/time-series | 800 | €0.144 |
GET /ai-search/overview/aggregated | 1,800 | €0.324 |
POST /ai-search/overview/leaderboard | 7,500 flat | €1.35 |
GET /ai-search/prompts-by-target | 200 per prompt returned | €0.036 per prompt |
GET /ai-search/prompts-by-brand | 200 per prompt returned | €0.036 per prompt |
Credit costs published in the SE Ranking Data API reference and wallet price on its API pricing page, checked on October 23, 2026. Euro figures are our division of €45 by 250,000 credits.
DataForSEO prices by task, with a queue to choose. Its LLM Scraper page publishes $0.0012 per result on the standard queue, $0.0024 on priority and $0.004 live, printed as "1,000 × 0.0012 = $1.20", with turnaround of up to 45 minutes, up to 5 minutes and up to 120 seconds on average. Its other families bill differently: LLM Responses at $0.0002 per task plus the token cost the model provider charges, AI Keyword Search Volume at $0.01 per task plus $0.0001 per row, and LLM Mentions at $0.1 per task plus $0.001 per row.
Put a real agency month against both. Ten clients, 30 prompts each, two engines, collected weekly, is 2,600 answers a month.
| A month for ten clients | What you get | Bill |
|---|---|---|
| API for GEO, 2,600 answers | answers plus share of voice, position, sentiment, ranked sources | $33.80 |
| DataForSEO, 2,600 scraper tasks | 2,600 fresh ChatGPT and Gemini answers with their sources | $3.12 |
| SE Ranking, 10 leaderboards, 2,000 prompts, 40 time series | a monthly snapshot per client, with a share of voice leaderboard | €91.26 |
Our arithmetic on the rates each vendor published, checked on October 23, 2026, under one stated basket: 2,600 answers for the two per-answer routes, and for SE Ranking one leaderboard call per client, 200 prompts returned per client and one time series per client per engine, which comes to 507,000 credits.
The three lines buy three different deliverables, so the ranking by price is not a ranking by value. The arithmetic settles only the shape of the invoice: DataForSEO scales with how much you collect, SE Ranking scales with how many reads you make over data it collected anyway, and the Enterprise rate SE Ranking publishes as "as low as $30/1M credits" moves its 200-credit prompt from €0.036 to $0.006, which is $6 per 1,000 rather than €36.
How you ask each one a question
On SE Ranking you name a domain, not a prompt. The base URL is https://api.seranking.com and the key goes in an Authorization: Token header:
curl -G https://api.seranking.com/v1/ai-search/prompts-by-target \
-H "Authorization: Token $SERANKING_API_KEY" \
-d target=example.com \
-d engine=chatgpt \
-d source=us \
-d scope=base_domain \
-d limit=100
The response carries a total, a date documented as the "date of the data snapshot", and a prompts array in which each row has the prompt text, an estimated monthly volume, a type describing how the target appeared, and an answer object holding text, the "full text of the snippet generated by the LLM", and links. The leaderboard endpoint is the one that computes: a POST with a primary target, one to ten competitors, a source country code and an engines array returns rows of rank, domain, share_of_voice, brand_presence, link_presence and is_primary_target.
On DataForSEO you send the prompt and wait for the engine. One POST carries up to 100 tasks, and results come back from task_get/advanced or through a postback_url:
curl -X POST https://api.dataforseo.com/v3/ai_optimization/chat_gpt/llm_scraper/task_post \
-u "$DFS_LOGIN:$DFS_PASSWORD" \
-H "Content-Type: application/json" \
-d '[{
"keyword": "best seo agency for saas in belgium",
"location_code": 2840,
"language_code": "en",
"force_web_search": true,
"postback_url": "https://example.com/hooks/dfs",
"postback_data": "advanced"
}]'
The task POST reference puts the prompt in keyword, up to 2,000 characters, with location_code or location_name and language_code or language_name required alongside it, priority switching to the faster queue at extra cost, and expand_citations available once force_web_search is on. Leave force_web_search off and you get the model answering from memory rather than from the web, which is a different measurement.
What comes back in the response
| Field an agency report needs | API for GEO | SE Ranking Data API | DataForSEO AI Optimization |
|---|---|---|---|
| Answer text | yes | answer.text, the snippet | markdown, plus a raw HTML endpoint |
| Cited URLs | ranked and sorted by category | answer.links | sources, with url, domain and title |
| Every page the model retrieved | not published | not published | search_results, duplicates and unused entries included |
| Brands named | entities[], aliases collapsed | brand_presence counts | brand_entities, ChatGPT scraper only |
| Position | avgPosition | average_position, with monthly history | not published |
| Share of voice | yes | leaderboard, against up to 10 competitors | not published |
| Sentiment | score, themes, per-brand perception | not published | not published |
| Fan-out questions | own endpoints | not published | fan_out_queries |
| Shopping items | yes | not published | chat_gpt_products, with price and merchants |
| Traffic estimate | not published | ai_traffic and ai_opportunity_traffic | not published |
| Prompt volume estimate | not published | volume per prompt | AI Keyword Data, a separate endpoint |
Field names published in each vendor's API reference, read on November 14, 2026. For API for GEO, the engines marked above as confirmed come from our hands-on test on October 26, 2026 rather than from apiforgeo.com.
Each side owns a column the other has no answer for. SE Ranking gives you prompt-level demand and a traffic estimate in the same row as the answer, which is the material for the slide where a client asks which questions are worth writing for. DataForSEO gives you search_results, described as "all web search outputs the model retrieved when looking up information, including duplicates and unused entries", which is the difference between what the engine read and what it chose to cite, and fan_out_queries, "related search queries derived from the main query". Our explainer on query fan-out in AI search covers why those sub-queries decide which pages end up cited.
Neither one returns a sentiment field, and only SE Ranking returns a share of voice. For what those metrics mean and how vendors differ on the denominator, our GEO metrics explainer sets out each formula.
Which engines each key reaches
| Engine, of the 10 we track | API for GEO | SE Ranking Data API | DataForSEO |
|---|---|---|---|
| ChatGPT | yes | chatgpt | AI Optimization, interface and official API |
| Gemini | yes, from our test | gemini | AI Optimization, interface and official API |
| Google AI Overviews | yes, from our test | ai-overview | SERP API, and the LLM Mentions database |
| Google AI Mode | yes | ai-mode | SERP API |
| Perplexity | yes | on the product page, not in the parameter examples | AI Optimization, official API |
| Claude | yes | no | AI Optimization, official API |
| Copilot | yes | no | no |
| Grok | yes | no | no |
| DeepSeek | yes, from our test | no | no |
| Meta AI | yes, from our test | no | no |
| Total | 10 | 5 | 6, across two keys |
Engine lists published on each vendor's product page and API reference, read on November 14, 2026. For API for GEO, ChatGPT, Perplexity, Google AI Mode, Claude, Copilot and Grok are listed on apiforgeo.com, and Gemini, Google AI Overviews, DeepSeek and Meta AI were confirmed in our hands-on test on October 26, 2026.
Two details worth carrying into a scoping call. SE Ranking's product page names Perplexity among the engines it covers, while the engine parameter in the AI Search reference is documented with four example values, ai-overview, chatgpt, gemini and ai-mode. Check the enum before you promise a Perplexity column. And DataForSEO's six engines need two products and two rate cards: Google AI Overviews and AI Mode come from the SERP API at $1.20 per 1,000 AI Mode tasks on the standard queue, which our SerpApi against DataForSEO comparison prices in detail.
Limits, plumbing and what the contracts allow
| SE Ranking Data API | DataForSEO AI Optimization | |
|---|---|---|
| Rate limit | 10 requests a second per key, 1 on trial accounts, 429 beyond | up to 10,000 calls a minute, 100 tasks per POST |
| Delivery | synchronous reads | Live and Task POST/GET, postback and pingback |
| SDKs | none published, Postman collection | Python, C#, TypeScript, Java |
| OpenAPI spec | not published | public repository |
| MCP server | https://api.seranking.com/mcp, OAuth 2.1 | hosted and self-hosted |
| Free entry | 14-day trial, 100,000 credits, no card | $1 trial credit and a free sandbox |
| Failed calls | "4xx and 5xx responses are not billed" | charged on task creation |
Published in each vendor's API reference and FAQ, checked on October 23, 2026.
DataForSEO is the better-equipped key for a pipeline, and it is not close. Webhooks, batch POSTs of 100 tasks, four official client libraries, an OpenAPI repository and a published ceiling of 10,000 calls a minute mean the integration work is scheduling rather than plumbing. SE Ranking's answer on that axis is its FAQ stating it "doesn't currently offer first-party SDKs", with a Postman collection instead, and an MCP server it describes as "centrally hosted and managed" while warning that "not every API endpoint is available through MCP".
The contract is where the ranking flips for anyone selling a dashboard. SE Ranking's API terms forbid, absent a prior written agreement, "reselling, sublicensing, or bulk redistribution of API data" and "embedding raw or substantially similar API metrics inside a paid SaaS product, including white-label dashboards". The same section then allows what most agencies actually do, stating that the customer may "provide clients with reports, summaries, analyses, recommendations". A monthly deck is inside the licence, a client login is not.
DataForSEO publishes no resale clause of any kind, which is quieter and not safer. Section 7.1 of its terms states that SERP data "shall not be used to compete with or adversely affect the business interests of the search engine providers", and silence on resale means a lawyer's reading rather than a vendor's permission. Our guide to a white label AI visibility API covers the routes that put resale in writing, and our ranking of AI visibility APIs for agencies weights that clause alongside price per client.
Mistakes we see on this comparison
- Comparing €36 to $1.20 as though they buy the same thing. One is a read from last month's snapshot, the other a fresh execution. Decide which your report needs, then compare inside that category.
- Promising a weekly chart on SE Ranking's Data API. The database refreshes monthly and every row carries its snapshot date, so three of four weekly points would repeat.
- Buying the Data API when you wanted your own prompts. The AI Search endpoints take a domain. Custom prompts are the AI Results Tracker in the Project API, on a subscription plus the add-on, or SE Visible.
- Forgetting the second DataForSEO key. ChatGPT and Gemini come from the AI Optimization API, Google AI Overviews and AI Mode from the SERP API, with separate endpoints and separate rates.
- Leaving
force_web_searchoff. Without it the model answers from training rather than from the live web, and the sources array thins out accordingly. - Selling a client dashboard on SE Ranking data. White-label dashboards are named in the forbidden list; reports, summaries and analyses are named in the allowed one.
Our view
Three defensible answers, depending on what the data has to do.
- Volume collection of fresh AI answers at the lowest rate. DataForSEO wins without an argument: $1.20 per 1,000 ChatGPT or Gemini answers on the standard queue, webhooks, 100 tasks per POST, four SDKs and an OpenAPI spec, with a $50 minimum payment and the entire computed layer left to your code. Our score: 7.7 of 10, held up by pricing and integration and held down by what the payload omits.
- A monthly picture of which prompts already mention a domain. SE Ranking's Data API is the cheaper route to something DataForSEO does not sell: prompt-level volume, a share of voice leaderboard against up to ten named competitors, and an AI traffic estimate, for any domain, without configuring a single prompt. You pay €36 per 1,000 prompts returned, accept a monthly refresh, and keep the data out of anything you resell. Our score: 4.3 of 10, pulled down by the integration column rather than by the data.
- Numbers per brand you can hand to a client without writing the analysis. API for GEO is our pick, at $13 per 1,000 answers down to $7 at its largest tier, pay as you go with no published minimum, all ten engines we track, share of voice, position, sentiment, themes, per-brand perception, fan-out, shopping and YouTube citations in one response, free reads over stored runs, resale allowed, and 9.0 of 10, the best score of the 55 APIs this site rates. Said plainly, its limits are a rate about ten times DataForSEO's per answer, and a documentation gap: four of its ten engines, along with the webhooks, the OpenAPI spec and the official SDKs, come from our test of October 26, 2026 rather than from apiforgeo.com.
For an agency the honest answer is often two keys rather than one, since DataForSEO collects and SE Ranking surveys, and they cost little enough together that the decision is about which report you sell. To try the third option, request a key on apiforgeo.com: it comes with a sandbox project. The reviews go further on each one, with SE Ranking Data API, DataForSEO AI Optimization and API for GEO scored criterion by criterion, and the duel API for GEO vs DataForSEO line by line.
Frequently asked questions
Is SE Ranking's API or DataForSEO cheaper for AI search data?
DataForSEO, on the published rates, by a factor of about thirty. Its LLM Scraper bills $0.0012 per ChatGPT or Gemini result on the standard queue, which is $1.20 per 1,000 answers, with a $50 minimum payment and no subscription. SE Ranking's AI Search endpoints bill 200 credits per prompt returned, and its €45 wallet buys 250,000 credits, so 1,000 prompts cost €36. The gap closes only at the Enterprise rate SE Ranking publishes as "as low as $30/1M credits", which puts the same 1,000 prompts at $6. Prices checked on October 23, 2026.
Can you track your own prompts with SE Ranking's Data API?
No. The AI Search endpoints take a target domain, an engine and a country code, and return prompts from SE Ranking's own database, which its documentation states is "refreshed on a monthly basis". Custom prompts live in a different product, the AI Results Tracker in SE Ranking's Project API, which runs daily checks on prompts you configure in a project and needs a Core, Growth or Enterprise subscription plus the API add-on at €39 a month billed annually. Reading the answers back from that tracker is free: "Retrieving prompt answers does not consume API credits or subscription quota".
Which AI engines does each API cover?
SE Ranking's AI Search documentation lists four values for its engine parameter, ai-overview, chatgpt, gemini and ai-mode, and its product page adds Perplexity, which the parameter examples do not show. DataForSEO needs two keys to reach six of the ten engines we track: the AI Optimization API scrapes ChatGPT and Gemini interfaces and calls ChatGPT, Claude, Gemini and Perplexity through their official APIs, while Google AI Overviews and AI Mode come from its SERP API at $1.20 per 1,000. Neither reaches Copilot, Grok, DeepSeek or Meta AI.
Does either API return share of voice or sentiment?
SE Ranking returns a share of voice and no sentiment: its leaderboard endpoint ranks your domain against up to ten competitors and returns share_of_voice, brand_presence and link_presence per domain, at 7,500 credits per call, or €1.35 on the wallet rate. DataForSEO returns neither, only a brand_entities item that "contains information on brands mentioned in the response". API for GEO, our pick after hands-on testing, returns share of voice, position, sentiment, ranked sources, themes and per-brand perception in one response at $13 per 1,000 answers, down to $7 at its largest tier, pay as you go with no published minimum.
Can an agency resell SE Ranking or DataForSEO data to clients?
SE Ranking draws the line in writing. Its API terms forbid, without prior written agreement, "Reselling, sublicensing, or bulk redistribution of API data" and "Embedding raw or substantially similar API metrics inside a paid SaaS product, including white-label dashboards", while explicitly allowing you to "provide clients with reports, summaries, analyses, recommendations". A client deck is fine, a client-facing dashboard is not. DataForSEO publishes no resale clause at all, and section 7.1 of its terms forbids using SERP data to compete with the search engines it came from.
How fresh is the data each API returns?
SE Ranking's AI Search database is "refreshed on a monthly basis", and each response carries a date field described as the "Date of the data snapshot", so a weekly report built on it repeats the same snapshot three times out of four. DataForSEO returns an answer collected when you ask, with published turnaround of up to 45 minutes on the standard queue, up to 5 minutes on priority and up to 120 seconds on average in Live mode. The two feed different reporting rhythms.